Automatic test data generation model by combining dataflow analysis with genetic algorithm

Dataflow techniques have been widely applied to test the object-oriented software, but the existing techniques mainly care how to arrange method sequence to test, not paying attention to how to generate test data automatically. Although the existing random test data generators can create test data automatically, yet information about the test requirement is not incorporated into the generation process, it may fail to find data to satisfy the requirements. This paper presents an automatic test data generation model that combine genetic algorithm (GA) with dataflow analysis. The model applies Control Flow Graph (CFG) which is partitioned into proper blocks to aid data flow analysis. Data Type Graph (DTG) presented in this paper facilitates the constructing of an object to realize automatic testing. Due to the search function of genetic algorithm, it improves the efficiency of test data generation. The experimental results prove our model to be more effective in test data generation and in coverage of dataflow than the random algorithm, showing the applicable significance of the model in software testing for both desktop and embedded systems.